hover-microinteraction-catalog

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill contains logic that ingests data from HTML data-* attributes and interpolates them into the Document Object Model (DOM).
  • In SKILL.md, the imagePeek function reads link.dataset.image and assigns it to an img.src property.
  • In reference.md, the split-hover recipe reads el.dataset.text and uses it to reconstruct the element's innerHTML by wrapping individual characters in <span> tags.
  • Ingestion points: HTML attributes data-text and data-image in SKILL.md and reference.md.
  • Boundary markers: None present.
  • Capability inventory: The skill is limited to DOM manipulation and does not have access to file system, network, or process execution tools.
  • Sanitization: No explicit sanitization is performed on the attribute values, though the character-by-character wrapping in split-hover naturally disrupts standard HTML tag injection.
  • [DYNAMIC_EXECUTION]: The skill utilizes JavaScript to dynamically create and modify UI elements at runtime.
  • In SKILL.md, imagePeek uses document.createElement and document.body.appendChild to create a fixed-position preview overlay.
  • In reference.md, the split-hover recipe uses innerHTML to dynamically generate a series of spans for staggering animations. These are standard UI development techniques for microinteractions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 05:48 PM
Security Audit — agent-trust-hub — hover-microinteraction-catalog